SCC-EAPS Special Lecture- Antonios Mamalakis (Colorado State University)

Explainable Artificial Intelligence to identify climate signals

During this presentation, we will demonstrate how Explainable Artificial Intelligence (XAI) can be used in an innovative prediction setting to identify robust climate signals from different forcing scenarios. Our approach will involve training a neural network to distinguish between simulated climates under different scenarios. We will then introduce and utilize the XAI method SHapley Additive exPlanations (SHAP) to identify the distinctive patterns that the network has learned. This analysis offers a purely data-driven investigation of climate signals.

For more information and Zoom password information, please contact Dave Wright: djwright@mit.edu